Creating Virtualization, and Separating Vocals......

Hello, Guys,

I’m an audiophile who has a tremendous love for jazz, not that it matters. I was wondering if anyone could tell me how to place most of the vocals into the left channel (or headphone), and place the instrumentals into the right channel (or headphone), and also how to virtualize audio for the 3D, reverb effect. Thanks for the help in advance. :smiley:

That’s usually done by grouping the existing vocal tracks into the left channel and grouping the existing instrumental tracks into the right channel during mixdown and post-production into the final show.

Unfortunately, the process doesn’t work in reverse. You can’t start with a mixed, finished show and derive the individual instruments and voices. It’s a frequently requested feature.

There is a rule of thumb about how much to delay one side of a two-track mono show with respect to the other to simulate a “deep” stereo effect. I don’t remember the numbers; I need to go look. You can experiment yourself with the Time Shift Tool. Use the menus under the little black arrow to the left of the blue waves > Split Stereo Track. Select one of the two tracks by clicking just above the MUTE button. Time Shift Tool (sideways black arrows) and push it later or earlier and listen to what happens. When you’re finished, Menu > Make Stereo Track and then from the top of the Audacity window, File > Export.

Audacity will not Save a standard sound file, you have to Export one.


Here it is.



Do you want to move existing vocals (dead Center) from an existing Stereo Sound such that they are on the left side (along with bass and drums)?
That’s possible if the Stereo was only created with panning.
Have you an sample to upload?

Here’s an mp3 that has been rotated with the following code in the Nyquist prompt.

(setf factor 0.707107)
(vector (sum (mult factor (aref s 0)) (mult factor (aref s 1)))
        (sum (mult (- factor) (aref s 0)) (mult factor (aref s 1))))

What I tried doing before was, I split the stereo track, and make the right channel mono, and keep the left channel in the original stereo format, then I also tried panning the sound to one side, but still didn’t produce the desired effect. Here are some samples of what I did below. But yes that’s exactly what I’m trying to do. I’ve never tried the Nyquist Prompt, that’s undiscovered territory for me. :laughing:

You can download the channel-mixer from the Nyquist effect WIKI Website, if you don’t want to work with the Nyquist prompt - and to have a permanent plug-in for this Task in your collection.
“Center-to-left” under presets does essentially the same as the code above (only with 50% instead of 71% Amplitude from each channel).
Your sample above is nearly mono in the first place.
Although the voice is being moved to the left side, all other instruments are equally so.
If I understand rightly, the track is already edited.
Try the code or the channel mixer with the original and it should work fine.

Here’s a link: Missing features - Audacity Support

Although it is a very nice recording of “Like someone in love”, I am afraid there’s not much to do with this particular Piece.
Blossom Dearie (if I do not err) has made 3 LPs with the song (about 1958) but I wasn’t able to find one that’s in proper Stereo.
Some titles by Blossom are already mixed as described - voice left / comping right. That was evidently in Fashion by some (re-)Producers.
However, LSIL is not amongst them.
It is at the Moment virtually impossible to seperate different instruments within a mono recording - science will Need some more years to accomplish this satisfactorily.
You could try to extract an instrumental Version with programs like Band-in-a-box.
But the results will be more than poor.
Regretfully yours

Thanks guys, for all your help. I’ve come to accept that I cannot do anything with this particular song, let alone, anything from the “Give Him The Ooh-La-La” album. Whoever at Verve Records was responsible for the re-mastering/restoration for the 1998 re-issue could’ve done a better job. Thanks again. guy. :smiley:

Sample 1.mp3 isn’t stereo, it’s “dual-mono” : the identical track on both left and right channels, (albeit at different levels).

If you want to add a (faux) stereo-feel , try Steve’s “pseudostereo” plugin

Wow! :open_mouth: That really did make a difference. It’s funny you mention the track is in a “dual-mono,” format–the album (1998 Re-issue) says recorded in stereophonic sound, I assume the technology to fully remaster/restore and create this effect wasn’t even around in 1998. This is exactly what I was looking for, thank you! :smiley: Thanka to all of you for your help!

i also found that you can create a stereo effect by delaying one channel in the change tempo effect. i used the lowest setting -.001, in a 300.15 second track, the delayed channel timed to 300.16

I am assuming to create a delay LFO/widening/chorus-y effect, one would do this then copy/past the same track underneath the new stereo-effected track, only in mono mix… you could even then EQ the mono channel and such for even more tweaking. This would also enable you to mix between the “wet” and “dry” effect, like in an FX send. I was wondering if anyone tried this yet. There is probably already a plug-in created that i haven’t used…
I am one of those headphone mixers, where the Haas effect sorta don’t happen. One effect OTB that i use frequently is an delay-LFO effect, where L+R channels have the basic delay of 20 or 30 ms. but there is an LFO also, which seems to compensate for just hearing two channels at different times thru headphones.
I suppose after stereo-izing one channel, you could apply a different phaser or pitch to each channel to create a chorusing effect. I have yet to experiment more with this and find the different effect combos that can mimic the chains i have been used to using with hardware setups.
I did find that i could not apply the phaser effect to only one channel, i got a “Nyquist” prompt (2.0.3 Windows using Vista).

You would probably not want to use the Change Tempo effect for this because that will cause one channel to gradually and progressively drift out of sync with the other channel. There is a “Delay” effect included in Audacity that can produce a constant amount of shift (that does not drift over time) Audacity Manual or you can create a delay using the Time Shift tool .

The optional Pseudo Stereo plug-in works by splitting each channel into multiple frequency bands, then delaying each band by different amounts. The mid-frequency band for left and right channels are kept synchronised so as to minimise unwanted stereo position shift caused by the Haas effect and the lowest frequency band is kept synchronised to avoid phase cancellation if played back through a sub-woofer system. The other frequency bands are time shifted by varying amounts in left and right channels independently to create the “stereo spread” effect.

Not without more information about exactly which effect you were using, with what settings on what tracks, but please start a new topic if you wish to discus that.
For Nyquist effects there is a separate forum board here: Nyquist - Audacity Forum

Now that I have changed the dual mono tracks using pseudo-stereo, is there away to separate the vocals to the left side and an instrument to the right using the Nyquist prompt? If I’ve asked this already, please pardon the redundancy.

If the original recording is mono, (or dual-mono), and the vocals and instrument occur simultaneously, and they overlap on the sound spectrum, then IMO the answer is no.

The pseudo-stereo separation is by frequency-content, not whether the sound is an instrument or vocal,
e.g. bass frequencies on slightly on the left , treble frequencies slightly on the left,
so under that system a bass singer will be put on the same (right) side of the stereo field as a bass guitar.

It is possible to pan the Sound according to its Amplitude (and frequency band for voice).
This works in the current example fairly good because the voice is the loudest Instrument.
You could Emphase the voice by simply attenuating the quieter sounds (with a noise gate or a compressor/Expander).
This gives the left channel. The right channel is made up with the difference of the original and the scaled left channel.
But this will only seperate voice/comping at different times, not when they are sounding simultaneaously.
There are professional Plug-Ins available that can separate the voice from Background within a mono file.
But I would wait 10 years until they are available as apps and for some 5 bucks or so - with the Advantage that the Quality has also increased to meet your Needs.

Here is a song I’ve been working on for the past few moments (minutes-hours) to create somewhat the desired effect. I uploaded a sample “My Old Flame,” by Dinah Washington (1955). I used the pseudo-stereo effect, and changed some the factors using the Nyquist prompt, and I also panned the audio at least 20% to the right. Any thoughts/comments? :smiley:

It seems that we have the same taste, at least in this genre.
The Piece is very quiet - is this intentionally?
The left channel has -14 dB, whereas the right has -23 dB.
A difference of about 15 to 18 dB would mean that you only hear the Sound from one Speaker.
If you want to pan the Sound 20%, the difference is much less than 7.7 dB.
Empiric Studies Show that People locate the Sound at 58 % (from the Center towards 100 % left/right) for 7.7 dB.
For 20 % it would be 2.26 dB (source: sengpiel Audio).
If you rfer to the desired effect, do you mean that the voice should appear more in the left channel and the rest rather in the right one?
At least, this has been the Goal up till now, if I am not mistaken.